When youth commit serious violent or sexual offenses, this often generates a call for more severe punishments and longer detention sentences. An important question is whether (long) detention sentences are effective in decreasing recidivism among serious young offenders. To estimate recidivism rates in serious young offenders and elucidate the link between sentencing (in terms of custodial vs. non-custodial and length of imprisonment) and recidivism, three multilevel meta-analyses were conducted. With a systematic literature search, 27 studies and four datasets were traced, involving N = 2,308 participants, yielding 90 effect sizes for overall recidivism, 24 for specifically violent recidivism, and 23 for the association between length of imprisonment and recidivism. The average weighted overall recidivism rate was 44.47% (95% confidence interval [CI]: 37.59-51.46%) over an average period of 8.68 years. The rate of violent recidivism was estimated at 30.49% (95% CI: 20.92-40.52%), over an average period of 11.45 years. Recidivism rates were higher when recidivism was defined as an arrest for any new offense rather than for a specific offense and in studies conducted in the United States versus European studies. Violent recidivism rates were higher in studies with longer follow-up periods. Based on the limited available studies, no difference in recidivism rates following custodial and non-custodial sentences were found, nor an association between length of imprisonment and recidivism. To increase rehabilitation chances for youth offenders, further research is warranted to better understand the impact of sentencing and to ascertain what is needed to make custodial and non-custodial sentences more effective.
Keywords: meta-analysis; recidivism; sentence; violent offenders; youth violence.